The proposed study will continue analyses of data collected on 1410 men and women aged 45 to 64 in Tecumseh, Michigan. The principal objective of these analyses is to improve the quality of dietary and health behavior data obtained from free-living populations in order to facilitate epidemiologic studies of the etiology of cancer and other chronic diseases. The data include current, retrospective, and historical information on dietary intake, health status, health events, and health behavior. Also there are data from 184 men and women who have provided surrogate dietary and health behavior information about their spouses. The aims of this proposal are to analyze: 1. agreement between subject-surrogate reports about diet and other health behaviors; 2. the interrelationships among dietary variables associated with dietary change from 1967-69 to 1982-83; 3. the reproducibility of long-term retrospective data on health status, health events, and health behavior; 4. and, the relationship of cigarette smoking and cessation of smoking to adiposity and dietary patterns. Statistical methods to be used include the kappa statistic, for assessing concordance; factor analysis techniques for defining dietary patterns; and analysis of covariance for analyzing relationships between smoking behavior and adiposity.